pugs.blogs.com | pugscode.org | pugs.kwiki.org | paste: sial.org/pbot/perl6 | <stevan> Moose... it's the new Camel ":P | .pmc == PPI source filters!
Set by Alias_ on 16 March 2006.
00:01 Quell joined 00:02 jisom joined 00:07 Quell joined
svnbot6 r9650 | fglock++ | pX/../PCR - compiler works again; AST is much cleaner; several emitter nodes need to be rewritten 00:17
TimToady theorbtwo: /<%hash1,%hash2>/ is not currently legal, but we do need *some* way of specifying that several grammatical categories should be treated in parallel, so that's a possible syntax, I suppose. Though it doesn't come out to an <expr> inside. 00:41
<%hash1|%hash2> seems more likely if we want it to track two hashes at once. 00:42
avar TimToady: I've always wanted to say this to you: "Thanks for Perl, I owe you a beer";) 01:33
TimToady yer welcome! 01:34
avar ;) 01:38
theorbtwo If it doesn't take general expressions, then using a syntax that doesn't make it look like a normal expression is probably a good idea. 01:47
avar Hell, screw promises of beers on IRC, I'll just donate some money;) Donating to the Perl Foundation is the preferred way to do that right? (not a question to TimT-d in particular)
avar not familar with the whole donation system 01:48
01:53 GabrielVieira joined
GabrielVieira hi 01:53
TimToady TPC works well for USians who want a tax credit. Other countries might or might not have other ways to optimize. 01:54
TPF, even.
avar fleh, a last name is a required field;/ 01:55
Can I donate by fixing the donation form?;) 01:58
It wants my last name and state/province, neither of which are applicable;)
GabrielVieira haha 02:02
02:08 Quell joined
svnbot6 r9651 | fglock++ | pX/../PCR - some tests pass 02:17
spinclad sure, donate to The Phone Company. you know you want to. (why, that's true, you do want to. listen to your inner voice. it will just feel so... right.) 02:21
jisom just hope they're not the next enron, of course if they are, they need the real money 02:22
02:23 taeli joined
svnbot6 r9652 | fglock++ | pX/../PCR - implemented backtracking control ':' 02:51
03:17 ayrnieu joined 04:15 avar joined 05:02 FurnaceBoy joined
Alias_ avar: You don't have a last name? 05:25
avar: Who are you, Madonna?
05:26 FurnaceBoy_ joined 05:45 PerlJam joined, qu1j0t3 joined 06:01 iblechbot joined 06:18 qu1j0t3 left 06:39 vborja joined
Khisanth Alias_: Prince! 06:43
07:09 nothingmuch joined 07:13 ayrnieu joined 07:26 pdcawley joined
nothingmuch feels proud: perlmonks.org/?node_id=537727 07:26
so much thought so early in the morning
szbalint If only Intrepid wouldn't litter the CB. :P 07:30
nothingmuch didn't notice 07:32
the reason I'm proud:
today I opened the door to the office with my left hand, instead of my right
i *locked* it and pulled on the key
and needed someone to tell me to turn it the other way 07:33
07:36 dduncan joined, Aankhen`` joined
Alias_ what's the word 07:36
Sounds like a great name for a piece of blog software 07:37
"We have a word for these kinds of musings in hebrew - it's the combination of the roots for philosophy and flatulence.
-nuffin"
nothingmuch hitpalzefut 07:43
he is mitpalzef
as opposed to philosophise:
mitp(h)alsef
and he is farting:
he is maflitz
svnbot6 r9653 | Darren_Duncan++ | r3220@darren-duncans-power-mac-g4: darrenduncan | 2006-03-18 23:40:30 -0800 07:46
r9653 | Darren_Duncan++ | ext/Rosetta/ : continued rewrite of Language.pod
07:57 dduncan left
Quell This is an /amsg. I am now going to sleep, although I suppose most of you won't care.. bye :( 08:14
nirgle heh 08:15
nothingmuch nothingmuch.woobling.org/cute_moose.jpg 08:17
08:33 pmurias joined
pmurias morning all, 08:33
08:40 Khisanth joined 08:57 nnunley joined 09:24 ghenry joined 09:30 kane_ joined
ingy rgs: ping 09:38
nothingmuch morning ingy 09:42
09:54 larsen joined 10:14 chris2 joined 10:26 bernhard joined 11:10 nothingmuch joined 11:16 KingDiamond joined
nothingmuch does Set.pm want a diet? web.engr.oregonstate.edu/~erwig/diet/ 11:53
12:03 ghenry joined 12:07 elmex joined 12:09 pmurias joined
pmurias does any one use VMS here? using AutoSplit for as a base for a per subroutine seperate compilation thingy and not sure if to remove the VMS handling bits? 12:31
ingy hi nothingmuch 12:32
ingy just wrote a subroutine to find and replace any heredoc content in a source with its sha1 digest 12:33
it was harder than I thought 12:34
nothingmuch hi ingy 12:40
didn't PPI help?
ingy I didn't look 12:41
nothingmuch wonders wtf happenned to his terminal font
ingy just realizes that he needs to fold pod blocks and heredocs at the same time 12:42
ingy writes a failing test 12:43
12:45 nothingmuch_ joined
ingy hi nothingmuches 12:46
nothingmuch err, sorry
ingy considers using PPI
nothingmuch abuses substances 12:49
it's the third time in years that I had more than one coffee a day
bah
evidently it's not enough
it's the *first* time in years...
13:00 nnunley_ joined 13:03 iblechbot joined 13:08 KingDiamond joined 13:11 knewt joined 13:19 chris2 joined 13:29 premshree_ joined 13:45 Limbic_Region joined 13:51 jisom joined
nothingmuch larsen: ping 14:06
larsen pong
nothingmuch larsen: got freenode /msg power? 14:07
larsen uh ? 14:08
nothingmuch as in registered with NickServ
appearantly you are =
=)
larsen tel 14:09
14:12 spo0nman joined
Arathorn right. 14:19
nothingmuch bangs head on table 14:24
14:32 mh_pia joined
spo0nman looks around. 14:33
jisom wonders what the map at the bottom of www.openbsd.org/goals.html would look like for parrot/pugs/perl 14:36
avar jisom: it's easy to make one
spo0nman is from India.
jisom yes 14:37
Arathorn mainly us, europe, .il & .tw i'd have thought
and .au
jisom have everyone record their longitude and latitude.....
avar jisom: no, easier than that
jisom maintainability? 14:38
avar www.frappr.com/mediawikidevelopers
Arathorn geolocate everyone's irc proxy's IP's? ;)
avar there's a web service made just for that
Arathorn ah, the magick of google maps
jisom but it's fun to reinvent the wheel!
14:39 mh_pia left
Limbic_Region Have you guys seen desert-island.dynodns.net:8000/perl...lanet.html ? 14:40
spo0nman Arathorn: many would have cloaks.
Limbic_Region It is a project of theorbtwo that places monks that are "online" on the glob
globe even
14:40 clydee joined
jisom when was it regenerated? 14:41
cause I'm not on it
jisom scratches his head thinking about what exactly he registered as 14:42
Arathorn that's cool :)
totally and unashamedly offtopic, but given not much else is happening and #perl6 is a fairly random sampling of techies: if you saw arasphere.net/device.gif in real life, would you think that it was a cellphone or a PDA (or both)? 14:44
jisom that ms's thing? 14:45
Arathorn no, something i did
jisom oh
newton?
Limbic_Region jisom - oh, I don't know how often it is refresshed
avar pda
Arathorn it's not based on anything - meant to be a random concept device
lumi PDA 14:46
Limbic_Region I think you need to show up in "Other Users" when it scans
avar A cell would have buttons for dialing stuff
spo0nman pda
Arathorn right, that sounds fairly conclusive - thanks for looking :)
avar and if those are made digitally with the screen + touch screen functionality it's a given that it does other stuff than dial people
Arathorn nods 14:47
Arathorn sets about adding more buttons
spo0nman wants a pda which has a twidler like keypad ... big display and a bluetooth headset to talk on.
basically T9 is good enough for me ... i can type fairly well with it ... i just want a bigger diplay on my mobile device. 14:48
Arathorn well, at least this thing'd have a decent sized display :) 14:49
spo0nman i like the 02 (seeo2.com) ... i havent bought it cuz it lacks the keypad and it runs windows X-(
jisom how do you ctrl-alt-del on a cell phone? 14:51
Arathorn you just powercycle it :|
Arathorn has a bunch of XDAs on his desk, and has learned to love it
spo0nman jisom: there's a powerbutton. the new nokia i have hangs fairly frequently.
jisom I guess they're pumping out cell phones faster than they can test them properly
Arathorn both symbian series 60 and windows mobile are fairly flakey 14:52
spo0nman in my opinion ... no one has got the mobile/pda combo right till now. everything iv seen just sucks. I was never a fan of PDA's anyway. 14:54
gaal 2 14:55
jisom if I remember right, apple's ancient newton got a lot of things better than current pda's, but just remember that pen and paper's older than the pda 14:56
Arathorn shame this thing's just a random concept design
Arathorn fantasises about creating his own handset/pda, running some nice hackable OS 14:57
jisom are there and pure bare posix systems? 15:02
spo0nman www.insearchofstupidity.com/Stupid_...ducts.html
jisom: the newton is featured.
jisom newton was just too expensive, too powerful for it's time
gaal that URL is quite... wordy 15:07
spo0nman :) he's funny. 15:08
15:26 GeJ joined
ingy hola 16:40
merlyn hey ingy 16:41
16:41 FurnaceBoy joined
merlyn started looking at Jemplate yesterday 16:41
ingy Jemplate kicks much ass 16:44
16:51 pdcawley joined 16:55 nnunley_ is now known as nnunley
nnunley ingy: Have you looked at trimpath? 16:56
ingy for js templating
?
nnunley nods. 16:57
ingy yeah, I like the Jemplate approach a lot better (*shock*)
nnunley grins.
ingy actually I really do, which is coincidental
Jemplate is a less complicated at the end of the day 16:58
once you compile your templates, they are just regular js files
browser cached, debuggable, fast
nnunley Ahh.
ingy it's all terribly simple from a JS standpoint. 16:59
and very complex from a TT standpoint 17:00
which is to say the TT is very rich, and Jemplate gets to support all of it
I would not want to port the TT compiler to pure JS, and as it turns out it wouldn't gain me anything to do so 17:01
17:01 elmex joined
nnunley nods. "Not much." 17:02
ingy I should PAR the whole thing, then no perl dependency 17:03
just a jemplate compiler 17:04
if I cared...
Supaplex googles a bit for this Jemplate phenomena 17:05
ingy simply (almost all of) TT2 for Javascript 17:06
merlyn it's in the cpan 17:07
I installed from there
the ajax stuff is cross-browser tested too?
Supaplex humm neat. 17:09
ingy is cpan.org b0rked?
merlyn what part of cpan.org? 17:10
DNS?
ingy nm
merlyn: load this in various browsers: search.cpan.org/src/INGY/Jemplate-0...index.html
the ajax stuff works whereever those tests pass 17:11
merlyn wow. it's running tests!
is that hitting your site?
... Files=10, Tests=44, 38.309 seconds 17:12
ingy merlyn: no, it's hitting your browser
;)
merlyn Uh. How are you testing ajax then? 17:13
that needs to hit a server
ingy what do you mean?
17:13 nothingmuch joined
ingy cpan.org is a server 17:13
merlyn but you don't have any CGI running there, do you?
ingy ajax don't need cgi
nnunley XMLHttpRequest can use a GET
merlyn ajax = xhtmlhttprequest()
Supaplex it's all static ;)
merlyn Oh... getting a static value 17:14
then you aren't necessarily testing params and stuff
nothingmuch møøse!
ingy sure I am
xerox nothingmuch: wow what does it mean?
merlyn mXXse. not sure what that's supposed to be.
actually I see \303\270\303\270
lumi &oslash;
merlyn ahh. moose. :) 17:15
nothingmuch xerox: it's monty python mooseness =)
xerox moo!
nothingmuch prepared a moose for cuteoverload.com 17:16
i hope they accept it
nothingmuch.woobling.org/cute_moose.jpg
xerox Oooh. 17:17
nothingmuch i also tried to register a domain today
møø.se
(/me is sleep deprived)
17:17 vel joined
xerox No unicode allowed in domain names, I fear. 17:17
nothingmuch BZZZZZZZZZZZZT
en.wikipedia.org/wiki/International...omain_name
otoh you are partly right
sweden's ICANN equivalent only allows swedish characters 17:18
i'm hoping their verification is teh b0rked
Arathorn i thought all the browsers have it turned off
nothingmuch b0rk b0rk b0rk
not all
safari has it on by default
Arathorn so that people don't register comedy non-printing-character unicode domains
which you then go phishing with
nothingmuch anywho
dinner time
we'll have mandatory unicode jokes later, if you like =) 17:19
Arathorn okay ;)
bon appetit..
nothingmuch the word is obligatory, right?
Arathorn either works 17:20
17:20 DesreveR joined
xerox Call now! Click SMILE FORALL . com! 17:21
ingy parses Perl code using pure Perl! 17:29
and they said it couldn't be done... 17:30
FurnaceBoy they said it would drive a man mad...
jisom no, they said only perl can parse perl, that doesn't mean you can't write a perl parser in perl :-p
merlyn but only if it's willing to execute some code to get it right 17:31
if you try to do a non-execute perl parser, you will get it wrong.
FurnaceBoy the cpu is willing but the ram is weak
17:32 ruz_ joined
merlyn ... www.perlmonks.org/?node=44722 17:32
avar a non-execute perl parser?
merlyn the "regex or divide" game is interesting there.
a perl parser that doesn't run BEGIN blocks (and by extension, "use") 17:33
you *must* run some perl code to properly *lex* later Perl code.
ingy I didn't say I parsed *all* of Perl ;)
merlyn and without a good lex, you don't get a solid parse.
ingy but I got 4 test cases passing
merlyn perl6 suffers the same way.
ingy and I feel good about it
and I also feel good for successfully trolling merlyn :p 17:34
merlyn because Larry was not willing to eliminate the regex/divide... er.. divide. :)
ingy merlyn: have you read the ecmascript spec 17:35
merlyn I've tried. :)
it's very ... boring.
ingy it basically has two grammers
merlyn great way to fall asleep.
ingy grammars
merlyn "my mommy has two grammas"
ingy one for inside a regexp and one for outside 17:36
merlyn that makes sense 17:37
nnunley ingy: Grammars with partial specialization.
jisom well regex and perl are two langauges.....rules and perl6 are two languages 17:38
merlyn Well, you could view the regex language as merely a submode of Perl 17:40
and there are even overlaps
for example, within a regex, you can have a perl var with its own syntax 17:41
jisom and perl in a regex
merlyn yeah, actually, thanks to (?{ .. })
and perl sorta "cheats" to parse that
but specifying it formally would be Very Hard. 17:42
17:42 SamB joined
jisom and then there's fun stuff like using a s/// inside the right hand side of a s/// :-D 17:42
ingy jisom: I just used that 17:43
in my Perl parser 17:44
jisom sometimes the only way to go, a little unreadable
only problem is you can only nest so far before you run out of available characters
ingy heh
FurnaceBoy damnIjustRanOutOfSpaces
nnunley You can go pretty far with Unicode. 17:45
ingy nnunley: which unicode chars can you use for regexp quoting?
m//s 17:46
m!!s
nnunley Hrm. I guess a real test is required.
jisom m{}s
m}}s
ingy møøs
XD 17:47
jisom m\07\07s
nnunley I was responding more to the concept of running out of characters, then actual practical application.
ingy nnunley: and I was puffing on your strawman
nnunley S'not made of straw. S'made of people.
jisom peopleman? 17:48
sounds like DID
avar I use ~ a lot 17:50
m~~, s~foo~bar~
It's pretty easy to type for me and I find that I need it a lot less than /
17:58 gilimanjaro joined
ingy nnunley: sorry I msged you on freenode and for some reason I can't identify 18:10
I didn't ignore you
nnunley No problem. 18:15
Are you not registered with nickserv? 18:16
ingy no 18:20
or I don't know
I hate freenode
nnunley Right. It forbids non-registered nicks from /msging 18:21
Supaplex it's a "feature"
*cough*
ingy now it worked
svnbot6 r9654 | fglock++ | pX/.../PCR - all tests pass 18:32
nnunley wonders what that commit fully translates to. 18:34
18:36 nirgle joined 18:38 ruz_ joined 18:39 fglock joined 18:55 szabgab joined 18:56 gilimanjaro joined
Arathorn pX/Common/Pugs-Compiler-Rules, i'd imagine 18:56
misc/pX/... even
fglock Arathorn: hi 18:58
Arathorn hia
18:59 szabgab left, szabgab joined
Arathorn hasn't had a chance to play with P-C-R since his EOL-munging escapes of the other night :( 18:59
fglock Arathorn: the new version has inheritable grammars, and some bugfixes - but it is a lot slower 19:01
I'm not sure if grammars should be inheritable at all - S05 says rules are subroutines (not methods?) 19:03
merlyn grammars are classes 19:04
therefore rules are methods
Supaplex depends on the flavor of cheese at the time.
fglock merlyn: hi!
merlyn all along, it's been the intention that to modify Perl, you merely "subclass" the Perl "grammar"
if that changed recently, I'd be surprised 19:05
hey fglock
fglock how about unnamed rules - which grammar they belong to? In perl5, would you have to create an unnamed grammar first? 19:06
19:08 larsen joined
fglock for example, what /\d+/ translates to, in p5? do{package _grammar; use base Grammar; _rule = ....; _grammar->_rule( @_ ) } 19:12
merlyn I think it matters only when you're looking at subrules 19:13
so if the rule never invoked subrules, no biggy
if the rule *did* invoke subrules, it'd use the current __PACKAGE__
but I'm not so sure about that last point
(or whatever __PACKAGE__ is called in perl6)
merlyn pops in his PT_BR training CDROM 19:15
fglock merlyn: :)
merlyn: there is a #perl-br channel, I think
szabgab ?eval (2|3) == (2|3) 19:17
should that return (true|true) ?
19:19 justatheory joined
fglock if the current package 'use base Grammar', then __PACKAGE__->rule(...) would just work (that's the unsugared version) 19:20
19:22 gilimanjaro joined
fglock szagab: that's (true|false), because 2 != 3 and 2==2 19:22
19:27 _meppl joined
jisom so what does 'if(true|false) {say "foo"} else {say "bar"}' do? 19:27
19:27 nirgle left
szabgab fglock: I am not sure I get why is this an explanation :) 19:27
avar | being a binary OR? 19:28
jisom | being any
szabgab on which side do you keep comparing with the 2 value and not with both ?
any of 2,3 == any of 2,3 should not be true true ?
merlyn in one universe where you run it, it returns true. in another, it returns false. :) 19:29
that's the $universe->fork function. :)
jisom quantum physics!!!!
any possibility that can happen does happen
merlyn except in no universe, do I end up with Jennifer Aniston. :( 19:30
fglock '|' starts threads - it executes (2==2), (2==3), (3==2) and collects the results into a new '|'
jisom but in infinitely many universes, all men look like Jennifer Aniston and all women look like Brad Pitt :-p
infinitely many != all 19:31
szabgab fglock: 3 comparisons ? 19:32
jisom four, he forgot 3==3 19:33
szabgab ok so he will get 2 trues and 2 falses so it gives me the avreage ?
fglock jisom: 'if' collapses the 'any' first, so 1|0 ends up being true, and 1&0 is false
jisom but that's no fun! 19:34
fglock szagab: the result is a 'set' - it works like hash keys (no repetitions)
merlyn why would all the women look like Brad Pitt? Ugh 19:36
I'm glad I don't live in any of those universes!
fglock merlyn: they will only collapse if you look at them 19:37
pasteling "fglock" at 200.102.181.35 pasted "named/unnamed rule perl5 API" (30 lines, 732B) at sial.org/pbot/16396 19:47
fglock does this look sane? 19:48
19:52 Muable joined
jisom yeah 19:53
pmurias rather nice 19:54
svnbot6 r9655 | szabgab++ | some simple strings related tests 19:56
pmurias does allowing a file to consist many compilation units sound like good workaround for the low lrep compile speed? 19:57
fglock ok - I'm finding out how to implement it :)
szabgab night
pmurias s/consist/& of/ 19:58
19:58 DesreveR is now known as r0nny
fglock pmurias: only if you are cacheing - but I think the grammar can be optimized 19:59
19:59 szabgab left
fglock pmurias: the new PCR is even slower, because of OO 20:00
pmurias how much can the PCR be sped up? 20:01
fglock pmurias: the grammar needs to use a dispatch table, instead of array-of-rules
20:01 rindolf joined 20:02 _meppl is now known as meppl
rindolf Møøse! 20:02
ingy Elk!
rindolf Ingy!
ingy Rindolf! 20:03
pmurias fglock: you mean <%control_statements>?
fglock pmurias: yes 20:04
pmurias how difficult/fun is <%hash> to implement? 20:05
FurnaceBoy <merlyn>why would all the women look like Brad Pitt? Ugh
women looking like Aniston are bad enough... 20:06
ingy FurnaceBoy: you're being a little harsh on Jen 20:07
fglock pmurias: see the <@xx> implementation - you need to implement the extra longest-match logic (not too difficult)
pmurias i could do it, but implemented that way it would be even slower 20:09
fglock pmurias: if you add it to lrep-compiler I'd migrate it to PCR
pmurias: you can cache the sorted keys
pmurias: you can cheat - use p5 regexes to match the keys 20:10
pmurias explain more please
20:11 ayrnieu joined
pmurias the <%hash> tries all keys and uses the one which matches the longest string or the one which is the longest regex? 20:13
fglock pmurias: anything you do in low-level perl5 would be faster than OO rules - even a foreach will do
pmurias: the keys are pre-ordered - it returns the first which matches (key eq substr) - keys are plain strings 20:14
pmurias: then it executes the hash value 20:15
20:15 gilimanjaro joined
fglock pmurias: <@array> needs to execute all matches, until one returns true 20:15
pmurias fglock: checked the S05 again, a bare %hash does it not <%hash> 20:17
in <%hash> the keys are rules not strings (sound complex to do fast ) 20:19
s/sound/&s/
fglock pmurias: right. we need %hash ! 20:20
jisom so you can finally use something other than a string as a key to a hash?
fglock jisom: yes - the actual internal 'key' is an id assigned to the object 20:21
jisom that can create some really interesting deeply nested structures
svnbot6 r9656 | fglock++ | pX/../PCR - unnamed rules are functions; named rules are methods 20:30
20:45 kanru2 joined 20:49 ruz joined 20:51 gilimanj1ro joined 21:12 DaGo joined 21:13 siosiosios joined, ycheng joined 21:14 FurnaceBoy_ joined
svnbot6 r9657 | fglock++ | pX/../PCR - added failing tests (subrule calls) 21:30
pmurias got to go to bed now,good night 21:49
21:52 Arathorn is now known as Aragone 22:04 samv_home joined 22:11 mugwump joined 22:14 Muable joined, mtk00 joined 22:16 stennie joined 22:21 vborja joined 22:23 penk joined
avar wth 22:28
a pseudoperl6 compiler in perl 6?;)
mugwump avar: yep. gcc is written in C, GHC is written in Haskell, ... 22:29
avar I mean the misc/pX/Common/lrep stuff 22:30
22:31 jiing joined, jiing_ joined 22:59 Limbic_Region joined 23:01 Quell joined
svnbot6 r9658 | fglock++ | pX/../PCR - low-level implementation of <...> 23:19
23:19 fglock joined 23:31 revdiablo joined 23:32 kisu joined 23:34 revdiablo joined
svnbot6 r9659 | fglock++ | pX/../PCR - named subrules work; lexical subrules don't 23:43
23:45 kanru2 is now known as kanru

Powered by Cro and the Raku® Programming Language.

Please report any issues / comments / feature requests as an issue on App::Raku::Log.

Thank you!